We have implemented a graphical user interface for the ADIS system, as a realization of the ADIS concepts and theories. The system as a whole lets the user discover, annote, publish, and execute composite Web services, and with this concrete implementation we have proved the concept of using ontologies in Web service matchmaking, as well as realizing automated Web service composition, and semi-automatic Web service execution.

Medical researchers today often work in a highly collaborative environment. The methods used to collaborate often include face to face meetings, unnecessary e-mail discussions, and huge amounts of documents sent and received in order to cooperate, share and contribute to each others work. The collaboration gets even more challenging as researchers today often do not work from the same location. For this study, the purpose has been to explore if there are ways in which the everyday work of dementia researchers can be supported and made more efficient by the use of web-based collaborative tools. A case study including four researchers related to dementia research has been conducted. A series of interviews with the researchers were done, and based on the findings from these, a prototype has been suggested and evaluated by the participants. The main findings shows that there is a need for further support during collaborative and coordinating activities in order to make the research process more effective. A web-based collaborative tool made with the medicine 2.0 approach in mind is suggested, which includes social networking, supporting collaboration between different actors, and openness and sharing between these actors. The main finding from this study is that there is a need for improved methods to support the collaboration and coordination between dementia researchers. Allowing the researchers to create social networks including sharing and collaboration related to files was seen as effective. It had potential of reducing the time spent revising documents as a result of tagging and notification functionality, in addition to improve coordination between colleagues because the awareness of what other project members were doing increased. The study also shows that there is a potential in using patient-produces sources such as blogs and illness communities in research, specially within qualitative research and social welfare studies. A need to improve the visualization of available data during the examination phase was discovered, however the need was not so strong as for improving collaborative activities.

I dette studiet ser vi på hvordan man kan praktisere brukersentrert utvikling i de tidlige fasene av systemutviklingsløpet for applikasjoner til Microsoft Surface.Vi har sett på to temaer knyttet til dette; introduksjon av ny teknologi for brukere, og metoder for low-fi prototyping og wizard-of-Oz brukbarhetstesting. Ved praktisering av brukersentrert utvikling er det ikke uvanlig at brukerne som involveres har en lav teknisk innsikt. Dette gjelder spesielt for nye teknologier som multi-touchbord, som fremdeles er lite utbredt. Viktigheten av å formidle mulighetene og begrensningene ved Surface til brukerne er derfor stor. I denne oppgaven blir det sett på hvilken merverdi domenespesifikke demonstrasjoner gir fremfor abstrakte når man skal formidle egenskapene ved Surface. Dette blir besvart gjennom to workshops med uavhengige grupper. Den ene gruppen får presentert abstrakte demonstrasjoner, mens den andre får presentert domenespesifikke. Ved å måle og sammenligne idérikdommen på de to workshopene blir det undersøkt hvilken av de to formidlingsmetodene som fungerer best. Resultatene viste imidlertid ingen signifikant forskjell i antall ideer som kom opp i de to workshopene, men typene av ideer hadde et merkbart skille. Det viser seg at domenespesifikke demonstrasjonseksempler ikke har merverdi overfor abstrakte, men at de to metodene snarere utfyller hverandre. I de tidlige fasene av systemutviklingsløpet til en applikasjon for Surface er det hensiktsmessig å utvikle en prototype som kan brukes i wizard-of-Oz brukbarhetstester, slik at brukerne kan evaluere designet og komme med forslag til endringer, nye krav eller bekrefte designløsninger. Vi ser i denne oppgaven på hvilke metoder, verktøy og materialer som bør benyttes ved prototyping, og hvordan disse kan brukbarhetstestes ved bruk av wizard-of-Oz på flere testbrukere samtidig. Ulike teknikker blir utprøvd og disse resulterer i en prototype som blir brukbarhetstestet for en aktuell brukergruppe. Resultatene fra brukbarhetstestene viser at papirprototyping og wizard-of-Oz flerbrukertesting av Surface-applikasjoner lar seg gjennomføre, men at det krever god planlegging. Blant viktige utfordringer som avdekkes er testpersonenes påvirkning av hverandre og viktigheten av å skille mellom fysiske og virtuelle GUI-elementer i prototypen.

The difference between Partition Testing and Random Testing has been thoroughlyinvestigated theoretically. In this thesis we present a practical study ofthe differences between random testing and partition testing. Thestudy is performed on the open-source project Buddi with JUnit and Randoop as test tools. The comparisonis made with respect to coverage rate and fault rate. The resultsare discussed and analyzed. The observed differences are statisticallysignificant at the 10% level with respect to coverage rate, in favour ofpartition testing, and not statistically significant at the 10% level withrespect to the fault rate.

In this thesis we have investigated how collaborative learning experiencescollected with mobile devices and represented in a timeline can help reflection.We have developed a prototype application on the Android platform in order tolook deeper into what kind of data that should be collected and how the data shouldbe visualized and interacted with. Today's mobile devices are powerful devices capable of capturing a variety ofinformation. Our investigation of related work and existing applications showedthat there are many applications developed for information collection. Todifferentiate from these applications, and support a central aspect of ourresearch question, reflection, we wanted the users to collaborate andperform actions on the collected data in order to promote reflection andenhancing learning experiences. We identified 5 high level requirements thatwe wanted the application to support; collection of experiences, being in theenvironment, representation in a timeline, collaborating on experiences andre-visiting experiences. All these requirements were achieved in the Timelineapplication. The functionality implemented was a result of taking advantage of the theoretical background, related work, evaluation and discussions with an expertin the area of reflection as well as co-operation with a EU project(MIRROR) thatare conducting research in the area of reflection.The resulting prototype was evaluated over 3 days in a real user setting duringa general assembly in MIRROR.From this evaluation, together with the work done in background theory andrelated work, we have analyzed our results and identified some characteristicsthat should be considered when developing a tool for enhancing reflection onlearning experiences in a collaborative environment.

In later years, the health care work in hospitals has become increasingly fragmented, in a sense where different people and professions are required for the treatment of every single patient. As a consequence, personnel should be assisted to greater awareness of what is happening, so that they can better plan where to put in their efforts. Making information about ongoing activities more accessible to its users is hence important, but this will in turn require increased distribution of sensitive data inside the hospital. The concept of flexible de-identification has been proposed as a solution for the privacy issues raised by this, but then again new issues emerge when it comes to how useful the de-identified data are to its authorized end users, in practice.A series of six rapid field tests was executed along with a literature review on de-identification. The purpose was to explore some ideas to how de-identification could be implemented for information screens located in public and semi-public hospital environments, such as hallways, where personnel are likely to see them. The appropriateness of several techniques for de-identification was hence evaluated for being used in real-time visualizations, in contrast to previous known applications of the concept. This input was in turn used to design a high-fidelity prototype for use in a series of four experiments in a usability laboratory. The experiments involved role-play sessions, where nurses from a university hospital used the prototype in a simulation of realistic ward work. In a focused interview directly afterwards, they each assessed the usefulness of having a system available in such locations, considering that the information was de-identified. Moreover, the nurses evaluated six alternative approaches to de-identification of the sensitive information, and ranked them with respect to which, if any, would be best suited for use in their regular work environment.The experiments indicate that users appreciate being notified via large screens when new information is available, but disagree on what is the preferred level of de-identification. Some would emphasize the legislative requirements and privacy issues raised, while others would put their own utility needs first. As a response to this, an interactive prototype was designed to demonstrate how users can be given interactive control over how identifiable the displayed information is. This idea of giving users flexible control over what is seen on a screen, depending on how they assess the context for access, is grounded in a framework for evaluation that considers the quality requirements of identification utility, legislation and usability.Useful applications of non-interactive de-identification to screens in public environments, are effectively disqualified by the legislative requirements regulating how personal health information can be disclosed. The de-identification can however be useful for enabling an intermediate security level, which can be accessed as long as there is a authorized user present. Appropriate techniques for achieving such de-identification, are found to be suppression of variables, coding, masking and generalization. With this overall approach, users may gradually authorize themselves until the required utility is reached, and hence be able to access useful information in public places. The information depth available must also be accordingly limited, so that the increased risk of abuse is mitigated. The result is possibly a security mechanism that is both legal to implement, it serves the utility needs of personnel, and it is more usable in practice than existing time-demanding login routines. Finally, these ideas have been included in the design of an interactive prototype, which still remains to see tested in practice.

The use of third party software components is increasing. By looking at developers at a Norwegian IT consultancy firm I find that developers are using components at an individual level and there is no leading agenda that promotes reuse. Developers that are used to finding and using components do so often and with few problems this practice is aiding and improving software development. However the frequency of use is not as high as it could be, either because of limitations of reuse or lack of knowledge and skill on how to find components. By encouraging the use of small software components and proposing simple guidelines on how to do so companies could increase both reuse and the benefits of them.

Denne oppgaven tar for seg småa- og mellomstore bedrifters (SMB) behov forBusiness Intelligence (BI), og utfordringer for slike bedrifter. Det diskutereshvordan disse bedriftene burde gåa fram, for å lykkes med BI. Det gjøresogsåa en vurdering på hvilke bedrifter som er egnet for å benytte seg avOpen Source BI-programvare.

Modern search engines have several data sources available to users, e.g. Newssearch, Image search and Video search. When a user enters a query in a searchengine, it is up to the user to choose a different source than the normal web search.On average, a user will only consider the first few occurrences in a search result anddo so in a few seconds. It would therefore be beneficial to the user experienceif the user did not have to limit the sources manually to refine a search.This project will evaluate different machine learning methods to classify relevantsources to a query. The goal of this is having an automated learning system thattakes some labeled input and uses this to help inform or direct the user to therelevant source.The project will take advantage of a Yahoo! product; Yahoo! Query LinguistAnalysis Service (abbreviated QLAS from now on and through the document). Thegoal is to incorporate semantic data from QLAS into the learning system. Thisshould augment the amount of information available to the learning system, andimprove its performance. It is not clear how this semantic data could be combinedwith the training data and incorporated in the learning system. A substantial partof the project will be to explore this.This project was done in cooperation with Yahoo! Technologies Norway AS (YTN).YTN develops Vespa, a search engine platform that has the possibility to searchfrom multiple sources. YTN is interested in researching the field of learning sourcerelevance to improve the search experience in Yahoo services. YTN is also interestedin researching ways data from QLAS could be used by Vespa to enable sourcerelevance classification when Vespa is used in a multiple-index setup.

I de siste årene har utviklinga innen IKT bidratt til at elektronisk samhandling mellom ulike aktører i samfunnet blir en stadig viktigere del av hverdagen. Denne trenden ser vi også i helsesektoren der store deler av arbeidshverdagen nå fokuseres rundt tett integrerte informasjonssystemer, slik som elektronisk pasientjournal. Disse journalsystemene utvikler seg fra å være et dokumentasjonsverktøy over til å inkludere stadig flere arbeidsprosesser. De blir dermed til et klinisk arbeidsverktøy som støtter de kliniske prosessene ved helseinstitusjonene.For å få til elektronisk samhandling er man avhengig av integrering, standardisering og tilpasning av informasjonsystemer hos de ulike aktørene. Denne oppgava diskuterer ulike strategier for standardisering ved utvikling av informasjonsystemer som skal skape en regional standard, samt hvilke arenaer og aktører som er viktige i denne prosessen. Oppgava er basert på et case studie av KMF-prosjektet (Kurve, Medisinering og Forordning) i Helse Midt-Norge, samt litteratur- og dokumentstudier. Det empiriske materialet er samlet gjennom observasjoner og intervjuer av ansatte ved pilotavdelinga, i tillegg til møter, telefonsamtaler og e-postkorrespondanse med ulike prosjektrepresentanter. Målet med denne oppgava er å synliggjøre hvordan ulike strategier, og det arbeidet som ulike aktører gjør underveis i stor grad påvirker utviklingsprosessen. Med tanke på at stadig flere kritiske og svært komplekse informasjonssystemer på tvers av ulike sektorer i samfunnet digitaliseres og samhandler, blir det enda viktigere å få en god forståelse av disse systemene og infrastrukturen de bygger opp. Denne oppgava forsøker derfor å dra fram noen utfordringer og aspekter som er viktige å tenke på ved utvikling av robuste systemer for samhandling.